Our study of the regional dimension of China's innovation system has in essence yielded five main findings. Firstly, contrary to what general wisdom may suggest at first sight, the regional concen-tration of R&D activities in China could not be found to be particularly stark. While they are indeed much stronger than the general disparities in economic development at the provincial level they more or less resemble those found in the former EU15 on a coun-try level. Furthermore, the current disparities among the countries of the EU27 typically exceed those among the provinces of China. As in most countries, regional disparities are lower for those expenditures more directly controlled by the state (e.g. by state owned firms or higher education institutions) and higher within the enterprise sector. The expenditures of the public research institutions, most prominently the CAS, which remain to a high degree concentrated on Beijing constitute an exception from this rule. Another aspect of this phenomenon is that the output of scientific activities (publications) is regionally more concentrated than the out-put of most technological activities (patents). Secondly, while the concentration of scientific activities does not vary much across fields, the concentration of technological activities does so quite substantially, more so than for example in Europe. Remarkably, moreover, it does so differently than in the European Union. While in Europe applications in the classic technological fields tend to be more concentrated, China shows a strong regional concentration in the high-tech fields (Correlation between Chinese and EU Pattern of Concentration by Field: -0.41). Thirdly, while the degree of regional concentration has diminished in all scientific fields, the picture is mixed with regard to technological activities. The strongest increases in regional concentration, in this context, can be attributed to the dominant role of Guang-dong in certain fields such as telecommunications. Remarkably, in contrast, there has been a substantial decrease in regional concentration with regard to biotechnology and microbiology and genetics. Fourthly, a juxtaposition of publications and patent applications yields an inhomogene-ous picture. In some fields, such as telecommunications, the centres of academic activ-ity are spatially quite separate from those of technological activity. In some others, such as Chemistry or Engineering, both activities are more synchronised. For the time being, however, there is little in our data that suggests strong linkages between academic and technological activities at the regional level. In contrast, the pattern of co-financing between public and private activity provides tentative evidence against it for many re-gions. Nonetheless the results suggest that the current regional pattern of distribution of scientific and technological activities in certain fields harbours potential to develop local linkages in the future. In other fields, substantial challenges continue to lie ahead. Finally, the pattern of technology-related trade is the most concentrated of all activities analysed in this study - far more concentrated than regional GDP or regional S&T ac-tivities. Moreover, it cannot be found to be highly co-incident with scientific or techno-logical activities on a regional level. Instead, it seems to reflect patterns of export proc-essing and assembly in e.g. the electronics and telecommunications sector. Thus, our findings are in line with the findings of Linden et al. (2007) as well as Koopman et al. (2008), quoted in the last OECD report (OECD, 2010). In summary, our regional analysis of the Chinese innovation system has painted the picture of a system in transition in which, despite substantial overall growth, disparities are not increasing but slowly decreasing - even though this decrease remains more evident for activities controlled by the state than for activities typically associated with the private enterprise sector. Moreover, it appears that while high-tech activities may still be concentrated on a number of key locations, activities in the medium-tech sectors are far more distributed. If there are strong linkages between scientific and technologi-cal activities at the regional level outside the prime locations, in contrast, remains ques-tionable. If anything, the evidence collected for this study cautions against easily taking such assumptions and suggests to individually check on the micro level. Nonetheless, the current pattern of scientific and technological specialisation implies that potentials are present for the future development of knowledge driven clusters at a regional.
